scripts/gen_registry_facts.py

#!/usr/bin/env python3
"""Harness: registry/seed_agents.yaml → router/registry_facts.pl (stdlib + PyYAML).

Not product code. Regenerates Prolog agent facts so symbolic_rules.pl stays
aligned with the seed registry. Prefer: scripts/check_prolog_registry_sync.sh
"""
from __future__ import annotations

import sys
from pathlib import Path

try:
    import yaml
except ImportError:
    yaml = None

ROOT = Path(__file__).resolve().parents[1]
SEED = ROOT / "registry" / "seed_agents.yaml"
OUT = ROOT / "router" / "registry_facts.pl"


def to_atom(agent_id: str) -> str:
    """Registry PascalCase id → Prolog atom (downcase first letter)."""
    if not agent_id:
        raise ValueError("empty agent_id")
    return agent_id[0].lower() + agent_id[1:]


def parse_seed_stdlib(text: str) -> list[dict]:
    """Minimal YAML subset parser for seed_agents.yaml (no PyYAML required)."""
    agents: list[dict] = []
    cur: dict | None = None
    list_key: str | None = None

    def flush():
        nonlocal cur
        if cur is not None:
            agents.append(cur)
            cur = None

    for raw in text.splitlines():
        if not raw.strip() or raw.lstrip().startswith("#"):
            continue
        if raw.startswith("- agent_id:"):
            flush()
            cur = {
                "agent_id": raw.split(":", 1)[1].strip(),
                "capabilities": [],
                "permissions": [],
                "dependencies": [],
                "control_plane": False,
            }
            list_key = None
            continue
        if cur is None:
            continue
        indent = len(raw) - len(raw.lstrip(" "))
        s = raw.strip()
        if indent == 2 and ":" in s and not s.startswith("-"):
            key, _, val = s.partition(":")
            key = key.strip()
            val = val.strip()
            list_key = None
            if key in ("capabilities", "permissions", "dependencies", "languages",
                       "input_types", "output_types", "required_evidence"):
                if val.startswith("[") and val.endswith("]"):
                    inner = val[1:-1].strip()
                    cur[key] = [x.strip() for x in inner.split(",") if x.strip()] if inner else []
                elif val == "":
                    list_key = key
                    cur[key] = []
                else:
                    cur[key] = [val]
            elif key == "control_plane":
                cur[key] = val.lower() in ("true", "yes", "1")
            elif key == "risk_class":
                cur[key] = val
            elif key == "domain":
                cur[key] = val
            else:
                cur[key] = val
        elif indent >= 4 and s.startswith("- ") and list_key:
            cur.setdefault(list_key, []).append(s[2:].strip())
    flush()
    return agents


def load_agents() -> list[dict]:
    text = SEED.read_text(encoding="utf-8")
    if yaml is not None:
        data = yaml.safe_load(text)
        if not isinstance(data, list):
            raise SystemExit("seed_agents.yaml: expected list")
        return data
    return parse_seed_stdlib(text)


def emit(agents: list[dict]) -> str:
    lines: list[str] = [
        "% SPDX-License-Identifier: AGPL-3.0-only",
        "% Copyright (C) 2026 David contributors",
        "% GENERATED by scripts/gen_registry_facts.py from registry/seed_agents.yaml",
        "% Do not edit by hand — regenerate via scripts/check_prolog_registry_sync.sh",
        "%",
        "% Agent atom map: registry agent_id → Prolog atom = downcase first letter",
        "%   LegacyCobol → legacyCobol,  Risk → risk,  Quant → quant",
        "% Predicates: agent_id/2, agent/1, control_plane/1, capability/2,",
        "%   permission/2, risk_class/2, depends/2",
        "",
    ]
    # agent_id/2 map first
    for a in agents:
        rid = a["agent_id"]
        atom = to_atom(rid)
        lines.append(f"agent_id('{rid}', {atom}).")
    lines.append("")
    for a in agents:
        atom = to_atom(a["agent_id"])
        lines.append(f"agent({atom}).")
    lines.append("")
    for a in agents:
        if a.get("control_plane"):
            lines.append(f"control_plane({to_atom(a['agent_id'])}).")
    lines.append("")
    for a in agents:
        atom = to_atom(a["agent_id"])
        for cap in a.get("capabilities") or []:
            lines.append(f"capability({atom}, {cap}).")
    lines.append("")
    for a in agents:
        atom = to_atom(a["agent_id"])
        for perm in a.get("permissions") or []:
            lines.append(f"permission({atom}, {perm}).")
    lines.append("")
    for a in agents:
        atom = to_atom(a["agent_id"])
        rc = a.get("risk_class")
        if not rc:
            raise SystemExit(f"missing risk_class for {a['agent_id']}")
        lines.append(f"risk_class({atom}, {rc}).")
    lines.append("")
    for a in agents:
        atom = to_atom(a["agent_id"])
        for dep in a.get("dependencies") or []:
            lines.append(f"depends({atom}, {to_atom(dep)}).")
    lines.append("")
    return "\n".join(lines)


def main() -> int:
    out = Path(sys.argv[1]) if len(sys.argv) > 1 else OUT
    agents = load_agents()
    if not agents:
        print("no agents loaded", file=sys.stderr)
        return 1
    text = emit(agents)
    out.parent.mkdir(parents=True, exist_ok=True)
    out.write_text(text, encoding="utf-8")
    print(f"wrote {out} ({len(agents)} agents)")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())